Research ArticleChemistry

Valence self-regulation of sulfur in nanoclusters

See allHide authors and affiliations

Science Advances  22 Nov 2019:
Vol. 5, no. 11, eaax7863
DOI: 10.1126/sciadv.aax7863


The valence self-regulation of sulfur from the “−2” valence state in thiols to the “−1” valence state in hydroxylated thiolates has been accomplished using the Pt1Ag28 nanocluster as a platform—the first time that the “−1” valent sulfur has been detected as S−1. Two previously unknown nanoclusters, Pt1Ag28(SR)20 and Pt1Ag28(SR)18(HO-SR)2 (where SR represents 2-adamantanethiol), have been synthesized and characterized—in the latter nanocluster, the presence of hydroxyl induces the valence regulation of two special S atoms from “−2” (in SR) to “−1” valence state in the HO-S(Ag)R. Because of the contrasting nature of the capping ligands in these two nanoclusters [i.e., only SR in Pt1Ag28(SR)20 or both SR- and HO-SR- in Pt1Ag28(SR)18(HO-SR)2], they exhibit differing shell architectures, even though their cores (Pt1Ag12) are in the same icosahedral configuration. Single-crystal x-ray diffraction analysis revealed their 1:1 cocrystallization, and mass spectrometry verified the presence of hydroxyls on Pt1Ag28(SR)18(HO-SR)2.


Metal nanoclusters are of great ongoing interest due to their atomically precise structures and intriguing properties (18). To date, several nanoclusters have been controllably synthesized and structurally determined (916). Because of their well-defined compositions and highly ordered surface structures, nanoclusters have served as models for investigating the surface coordination chemistry of nanoparticles at the atomic level (1724), and thiolated metal nanoclusters have been used to study the coordination modes of metal-sulfur interactions (1, 1720).

Most thiolated nanoclusters bear their sulfur atoms on their surface in the form of staple-like motifs, from monomeric M(SR)2 to heptameric M7(SR)8 (1, 20, 25), that protect and stabilize the surface of the nanocluster kernel. In particular, some bare S atoms (without a linking carbon tail) have been observed on the outermost surface or in the kernel of nanoclusters even though these bare S-containing nanoclusters are prepared in the absence of bare S sources (2628). The bare nature of these S atoms is believed to arise from the propensity of these nanoclusters to adopt the most thermodynamically stable structures (1, 2628). However, regardless of the nature of the sulfur atom (as bare S, or part of a larger motif), all are present in the “−2” valence state, the same as that of their parent thiols (1, 19, 2628).

The very narrow range of the valence state of sulfur in cluster science does not reflect the broad range of valence states of sulfur in nature, where −2, −1, 0, +2, +4, and +6 are all known (2933). Moreover, the ability of sulfur to self-regulate its valence is of great significance in both life sciences and materials science, being implicated in a multitude of vitally and economically important processes including the energy conversion of thiophilic bacteria and in Li-S batteries (2935).

Generally, the valence regulation of sulfur is a two-electron (2e) transfer process: The valence state of sulfur alters from −2 to 0 upon its oxidation from R-S-H to R-S-OH, and then to +2 in R-S-(O)OH and +4 in R-S-(O)2OH (29, 30, 3537). Sulfur atoms of uneven valence state (i.e., S−1) are rare and have only ever been observed in disulfides ([-S-S-]2−; e.g., FeS2, Ph2S2, and cysteine) (30, 31, 3638), although it is currently unknown whether a sole-sulfur atom with a “−1” valence state (S−1) experimentally exists—the “−1” valence state in the above examples is actually averaged over two sulfur atoms. However, it has been hypothesized that metal nanoparticles that catalyze the oxidation of thiols (e.g., from thiol with “−2” valent sulfur to sulfenic acid with “0” valent sulfur) deliver two electrons sequentially (3942), implying a possible existence of an intermediate containing a sole sulfur with a “−1” valence state. The discovery of such a sulfur atom should yield insights into the valence regulation processes of sulfur and their mechanisms.

Here, we report the self-regulation of the valence state of sulfur from the “−2” valence state in thiols to the “−1” valence state in hydroxylated thiolates on the basis of a pair of cocrystallized nanoclusters, formulated as Pt1Ag28(S-2-Adm)20 (Pt1Ag28−1; S–2-Adm = 2-adamantanethiol) and Pt1Ag28(S-2-Adm)18(HO-S-2-Adm)2 (Pt1Ag28−2). The “−1” valent sulfur has been observed as a sole sulfide (i.e., S−1), using Pt1Ag28−2 as a platform. The two sulfur atoms linking the “-OH” groups in Pt1Ag28−2 represent the valence self-regulation of sulfur from a “−2” valence state (in parent thiols) to a “−1” valence state. Structurally, both Pt1Ag28 nanoclusters are composed of a Pt1Ag12 kernel with an icosahedral configuration; however, the staple motifs of these two nanoclusters are entirely different. Single-crystal x-ray diffraction (SC-XRD) analysis revealed their 1:1 cocrystallization in unit cells. Electrospray ionization mass spectrometry (ESI-MS) results confirmed the cocrystallization behavior, neutral state, and 8e free electron count of each Pt1Ag28 nanocluster.


Pt1Ag28−1 and Pt1Ag28−2 were synthesized concomitantly via a one-pot synthetic method (see Materials and Methods for more details). Optical absorption of crude products showed two prominent peaks at 460 and 560 nm and two shoulder peaks at 360 and 435 nm. The crude and pure (crystal) products exhibited similar ultraviolet-visible (UV-vis) spectra, reflecting the high yield of the synthesis (fig. S1A). Thermogravimetric analysis (TGA) was performed to verify the purity of the synthesized nanoclusters; a weight loss of 51.45% (fig. S1B) was observed, which is consistent with the calculated loss of 50.98% for Pt1Ag28−1 and 51.25% for Pt1Ag28−2. X-ray photoelectron spectroscopy (XPS) and inductively coupled plasma (ICP) measurements validated the atomic ratio of Pt/Ag in Pt1Ag28−1 and Pt1Ag28−2 (fig. S1, C and D, and table S1); the results were in good agreement with the theoretical value (1/28). Infrared (IR) characterization of clusters was carried out, and the signal of 3297 cm−1 in IR demonstrated the presence of the hydroxyl group in clusters (fig. S1E). 1H nuclear magnetic resonance (NMR) of Pt1Ag28 clusters was performed (fig. S1F); however, the H from the hydroxyl group is hard to detect for two reasons: (i) Compared with the H from thiol ligands, the H from the hydroxyl ligands are fewer; specifically, there are 600 H atoms from the thiol ligands in the Pt1Ag28 cluster system, whereas only two H atoms exist from the hydroxyl. (ii) The H atoms on the hydroxyl are much more active than those from the thiol.

Structurally, the icosahedral Pt1Ag12 kernel is observed in both Pt1Ag28−1 and Pt1Ag28−2 (Fig. 1A), but the composition and arrangement of their capped motifs are totally different (see table S2 for more details of bond length comparison between two Pt1Ag28 clusters). Pt1Ag28−1 contains a Pt1Ag12 kernel wrapped by an Ag10(SR)12 ring motif (where SR = S-2-Adm) to give a structure of the formula Pt1Ag22(SR)12 (Fig. 1B), which is further protected by four staple-like motifs (two SR-Ag-SR-Ag-SR dimers and two Ag-SR monomers) to form the complete Pt1Ag28−1 (Fig. 1C). Pt1Ag28−2 contains an icosahedral Pt1Ag12 kernel that is the same as that in Pt1Ag28−1 (Fig. 1A), wrapped by an Ag8(SR)10 ring motif, which is much smaller but more regular than Ag10(SR)12 in Pt1Ag28−1 (Fig. 1D). Furthermore, the obtained Pt1Ag20(SR)10 [Pt1Ag12 + Ag8(SR)10] architecture is stabilized by two SR-Ag-SR-Ag-SR and two Ag-SR-Ag-(HO-SR) motifs, giving rise to the complete Pt1Ag28−2 (Fig. 1E). The length between S and O in the HO-S(Ag)-R unit is determined as 1.81 Å, which further verifies the single bond between S and O (notably, the general bond length of S═O ranges from 1.40 to 1.45 Å). No counterion was found in the crystal structure, implying a zero valence state for both Pt1Ag28−1 and Pt1Ag28−2, and the nominal electron counts of both nanoclusters are 8e (28-20 for Pt1Ag28−1; 28-18-2 for Pt1Ag28−2). Similar to other superatom clusters with electron shell closures (1, 2, 4), the eight free electrons of both Pt1Ag28 clusters are assigned to the M13 kernels, and thus, the shell metals tend to present an oxidized state (i.e., Ag+).

Fig. 1 Structures of Pt1Ag28 nanoclusters.

Structural anatomies of Pt1Ag28−1 (A to C) and Pt1Ag28-2 (A, D, and E). Color codes: green sphere, Pt; violet sphere, Ag in the icosahedral M13 kernel; blue sphere, Ag in the motif structures; yellow sphere, S; red sphere, O; gray sphere, C. For clarity, the hydrogen atoms are not shown.

Notably, there are two thiols bonding with hydroxyls in Pt1Ag28−2, giving rise to the “HO-SR-” units, the S atom of which is in the “−1” valence state [i.e., “−1”(integral) − “−1”(HO-) − “1”(R) = “−1”(S)]. The “−2” valence state of the S from the thiol ligand has been regulated to the “−1” valence state in the “M-SR-OH” unit. To the best of our knowledge, this is the first time that a single sulfur atom has been observed to be in a “−1” valence state (S−1). Besides, although sulfur is known for its ability to undergo valence tautomerization (2931), the valence state of sulfur remains unchanged in this work (i.e., “−2” valent sulfur in thiolates and “−1” valent sulfur in hydroxylated thiolates) because of the immobilization effect of the metallic kernel to the surface structures of clusters. The S−1 in the HO-S(Ag)-R unit contains an unpaired electron and is electron paramagnetic resonance (EPR) active. The EPR spectra of these nanoclusters are shown in fig. S1G. The EPR spectrum of the clusters (in CH2Cl2, at 104 K) showed an S = 1/2 signal with g = (2.0193, 2.0112, 2.0033) [g = (2.0178, 2.0070, 2.0000) at 4 K], unambiguously identifying the paramagnetism of the S−1.

The observation of S−1 has many implications, one of which is the sulfur oxidation process catalyzed by metal nanoparticles (3942). The oxidation of sulfur from the thiol to the sulfonic acid has been previously investigated (Fig. 2) and found to entail regulation by sulfur of its valence state from “−2” in thiol (R-S-H) to “0” in sulfenic acid (R-S-OH), then to “+2” in sulfinic acid [R-S-(O)OH], and finally to “+4” in sulfonic acid [R-S-(O)2OH] (3638). An intermediate of the form R-S(M)-OH (i.e., S−1) has been hypothesized to act as an intermediate in the oxidation from thiol to sulfenic acid, but a detailed, mechanistic understanding of the precise role of the nanoparticle catalyst remains a mystery. Our work provides new insight into the metal catalysis from thiol to sulfenic acid—the oxidation from thiol to sulfenic acid may not be a direct 2e transfer process but a stepwise 1e-1e electron transfer process [from R-S-H to R(Ag)-S-OH, then to R-S-OH; Fig. 2]. The observation of hydroxylated thiol ligands on Pt1Ag28−2 is of great significance not only because the “−1” valent sulfur fills the vacancy of thiol-oxidation pathway but also because the R(Ag)-S-OH acts as a possible intermediate among the nanoparticle-catalyzed sulfur oxidation. Because of this intermediate state, the S─O bond length (1.81 Å) in the R(Ag)-S-OH unit is longer than the general S─O bond length that ranges from 1.50 to 1.70 Å.

Fig. 2 “−1” valent sulfur in hydroxylated thiolates.

Sulfur-oxidation pathway from thiol to sulfonic acid. The “−1” valent sulfur in Pt1Ag28−2 supports the hypothesis that there is an intermediate between the “−2” valent S in thiols (R-SH) and the “0” valent S in sulfenic acids (R-S-OH).

Another significant implication directs to the valence state regulation of sulfur in biological systems (30, 34, 36). The valence self-regulation of sulfur in cellular proteins (e.g., the cysteine redox cycle) continues to be at the center of an extensive array of sulfur-based biological science (30, 34, 36). Several of these valence self-regulations are believed to be dependent on the electron transfer process of metal-sulfur clusters. For example, in most iron-sulfur proteins, the clusters function as electron-transfer groups in mediating the one-electron transfer. As to these one-electron transfer processes, our reported metal-S(OH)-R unit may also exist as a special state in valence regulations of sulfur in biological systems.

SC-XRD also revealed that Pt1Ag28−1 and Pt1Ag28−2 cocrystallized in a single cell unit—a phenomenon that has also been observed with (AuAg)45(SR)27(PPh3)6 and (AuAg)267(SR)80, and Ag40(SR)24(PPh3)8 and Ag46(SR)24(PPh3)8 (43, 44). As with both these pairs of nanoclusters, the occupancy of each Pt1Ag28 nanocluster in a unit cell is exactly 50% (fig. S2). In unit cells of Pt1Ag28−1 and Pt1Ag28−2, the lamellar cocrystallization is observed (Fig. 3, A and B), and the interlayer distance is 33.978 Å (calculated from the gap between each Pt plane). Akin to the face-centered cubic (fcc) nanoparticle lattice, the lamellar cocrystallized nanoclusters are organized into a square packing mode within the {100} plane (Fig. 3C). However, perpendicular to the x direction, the nanocluster lattices stack up obliquely, which is different from the arrangement in the fcc lattice (Fig. 3D).

Fig. 3 Cocrystallization of nanoclusters.

(A and B) Cocrystallization of Pt1Ag28−1 and Pt1Ag28−2 in packing mode. The interlayer distance is calculated as 33.978 Å. Views from (C) x direction, (D) y direction, and (E) z direction. Color codes: violet sphere, Ag in Pt1Ag28−1; blue sphere, Ag in Pt1Ag28−2; green sphere, Pt; yellow sphere, S; red sphere, O. For clarity, the carbon and hydrogen atoms are not shown.

ESI time-of-flight mass spectrometry (ESI-TOF-MS) analyses were performed to verify the coexistence of Pt1Ag28−1 and Pt1Ag28−2 in the cluster sample and confirm the presence of the hydroxyl in Pt1Ag28−2. As shown in Fig. 4A, two peaks were detected in the mass spectrum of the Pt1Ag28 sample, demonstrating the cocrystallization of these nanoclusters. Furthermore, upon magnification, the mass signals can be seen to show a characteristic isotopic pattern with peaks separated by m/z (mass/charge ratio) of 1 Da (in the positive mode)—both nanoclusters exhibited an overall +1 valence state. The mass spectrum was also a perfect match for the calculated isotopic distribution of both [Pt1Ag28−1 + Cs]+ and [Pt1Ag28−2 + Cs]+ (centered at 6694.01 and 6728.01 Da, respectively), implying the zero valence state of these two nanoclusters. The mass gap between these two peaks was calculated as 34 Da, which was exactly equal to the 2*MOH difference compared with the compositions between Pt1Ag28−1 and Pt1Ag28−2.

Fig. 4 ESI-MS results of nanoclusters.

(A) ESI-MS result of the Pt1Ag28 sample (including Pt1Ag28−1 and Pt1Ag28−2) and the comparison of the experimental and calculated isotopic patterns. (B) ESI-MS result of the deuterated Pt1Ag28 nanoclusters (including Pt1Ag28−1 and Pt1Ag28−2-D) and the comparison of the experimental and calculated isotopic patterns. (C) Comparison of ESI-MS results of Pt1Ag28-H (including Pt1Ag28−1 and Pt1Ag28−2) and Pt1Ag28-D (including Pt1Ag28−1 and Pt1Ag28−2-D). Note that (i) the mass peaks around 6694.01 Da remain (highlighted in gray) because no hydroxyl is bonded on Pt1Ag28−1, and (ii) the 2 Da of the mass gap between the peaks of 6728.01 and 6730.01 Da is precisely the double of the mass peak between H and D (highlighted in orange).

To further confirm the existence of the hydroxyl group on the Pt1Ag28−2 surface, the H2O, CH3OH, and NaBH4 reagents used to synthesize the Pt1Ag28 nanoclusters were substituted by D2O, CD3OD, and NaBD4, respectively, and the product nanoclusters were submitted for ESI-MS. Two discrete peaks were detected in the ESI-MS spectrum (Fig. 4B), corresponding to Pt1Ag28−1 and Pt1Ag28(S-2-Adm)18(DO-S-2-Adm)2 (Pt1Ag28−2-D for short); the mass gap between them was calculated to be 36 Da, exactly equal to the 2*MOD difference between the corresponding nanoclusters. Figure 4C depicts the mass spectra of Pt1Ag28-H and Pt1Ag28-D between 6670 and 6750 Da. Specifically, the mass peaks around 6694.01 Da that correspond to Pt1Ag28−1 remain because no hydroxyl exists on the surface of Pt1Ag28−1. However, another peak (around 6728.01 Da) reflects a 2-Da (6730.01 to 6728.01 Da) increase after the hydrogen sources are deuterated, a mass gap equal to double of M(D-H) [i.e., 2(MDMH)]. More mass details of Pt1Ag28-D are depicted in fig. S3. Furthermore, the H2O in the synthetic procedure was replaced by H218O; ESI-MS spectrum of the product still presented two discrete peaks, Pt1Ag28−1 and Pt1Ag28(S-2-Adm)18(H18O-S-2-Adm)2, at 6694.01 and 6732.01 Da, respectively (fig. S4). The mass gap between these two peaks is 38 Da, which is exactly double the mass of H18O (fig. S4). Together, these results unambiguously confirm the existence of the hydroxyl on the structure of Pt1Ag28−2.

To shed light on the overall reaction process from Ag-Pt-SR-PPh3 complexes to Pt1Ag28 clusters, we monitored time-dependent ESI-MS spectra. By systematically investigating the 43 species in the mass spectra (fig. S5), we found that the size growth from metal complexes to Pt1Ag28 clusters involved two stages: (i) from Pt-Ag-SR-PPh3 metal complexes to PtAg-based intermediate clusters [including Pt1Ag14(SR)102+, Pt1Ag22(SR)17+, and Pt1Ag28(SR)18(PPh3)4 clusters] and (ii) from intermediate PtAg-based clusters to stable Pt1Ag28−1 and Pt1Ag28−2 clusters. Besides, the Pt1Ag28−1 and Pt1Ag28−2 nanoclusters appeared concurrently (after 1 hour of the reduction of NaBH4). Although the ratio between two Pt1Ag28 clusters changes a little in the synthetic process, the ratio between them remains as approximately 1:1 (fig. S5C).

On the basis of the ESI-MS spectra (fig. S5), the −1-valent sulfur [or HO-S(Ag)-R unit] was only observed in the Pt1Ag28−2 cluster but was not found in both intermediate species and nanocluster fragments. Two channels of the sulfur hydroxylation are proposed here: (i) The reaction between Pt4+ and R-S-H gives rise to Pt2+ and R-S-O (see fig. S5E), wherein the valence state of sulfur changes from −2 to 0; then, the presence of BH4 reduces the 0-valent sulfur to the −1-valent sulfur, and the obtained −1-valent sulfur is further captured by the Pt1Ag28 structure to produce the Pt1Ag28−2 nanocluster. However, the presence of −1-valent sulfur in other species is hard to detect probably because of the instability of them [that is, the Pt1Ag28 kernel is able to immobilize and stabilize the HO-S(Ag)-R structure]. (ii) The R-S-OH (still from the oxidation of R-S-H by Pt4+; see fig. S5E) is directly captured by the Pt1Ag28 structure and catalyzed by the cluster@BH4 system at the same time to produce the HO-S(Ag)-R unit. Other hydroxylation approaches may happen; however, we cannot determine the precise hydroxylation process by the current testing methods. Future works will focus on the precise hydroxylation process on the Pt1Ag28−2 nanocluster.

One noteworthy observation was that although the Pt1Ag28 nanoclusters would not appear in the absence of PPh3, the PPh3 ligand is not present in the final products. We proposed that the PPh3 ligand might act as a reservoir of metal ions, reminiscent of the function of bidentate phosphine in the synthesis of Ag74 (45). We then performed 31P NMR experiments at each stage of the synthesis to track the fate of the PPh3. As shown in fig. S6A, PPh3 that dissolved in CD2Cl2 showed one peak at −5.61 parts per million (ppm), representing the “free” phosphine ligand. Upon addition of AgNO3, the P signal was shifted to 8.38 ppm (fig. S6A, red)—a significant change caused by the coordination between PPh3 and AgNO3. However, no such marked change in chemical shift was observed upon the addition of H2PtCl6 to the solution of Ag-PPh3 in CD2Cl2 (fig. S6A, blue). Furthermore, the addition of thiol ligands to the above Pt&Ag-PPh3 solution further shifted the P signal to 8.97 ppm, in the absence or presence of the Pt source (fig. S6A, green and purple), which suggested the formation of Ag-PPh3-SR complexes. Moreover, no P signal was found as to the final nanocluster product (fig. S6A, brown), which is in accordance with the structures of Pt1Ag28−1 and Pt1Ag28−2 (with no PPh3 ligands). Therefore, PPh3 serves as a reservoir of Ag in this synthetic procedure.

Pt1Ag28(S-1-Adm)18(PPh3)4 was obtained by substituting 2-AdmSH used in the synthetic procedure by 1-AdmSH (figs. S6, B and C, and S7, A and B) (46). Figure S6C depicts the 31P NMR spectrum of the S-1-Adm protected Pt1Ag28; from the peak at 26.05 ppm, Ag-P coordination between the Ag-PPh3 units on the Pt1Ag28(S-1-Adm)18(PPh3)4 nanocluster can be inferred. In this context, the nature of the generated cluster [i.e., (i) Ag-S(R)-OH units in S-2-Adm protected Pt1Ag28 or (ii) Ag-PPh3 units in S-1-Adm protected Pt1Ag28] depends on the differences between HS-2-Adm and HS-1-Adm ligands. Two factors are thought to account for the differences: steric hindrance and the existence of α-H (presented in HS-1-Adm but not in 2-S-Adm). To determine which of these was responsible for the formation of Ag-S(R)-OH, we prepared a new Pt1Ag28 nanocluster with cyclohexyl mercaptan, which has a similar α-H as HS-2-Adm. ESI-MS results demonstrated the presence of Ag-PPh3 but the absence of Ag-S(R)-OH on this newly obtained Pt1Ag28 nanocluster, the chemical formula of which was determined to be Pt1Ag28(S-c-C6H11)18(PPh3)4, the same as the Pt1Ag28 nanocluster capped by HS-1-Adm (with no α-H) (fig. S7). Consequently, the influence of the α-H effect on the formation of Ag-S(R)-OH was ruled out, and thus, we propose that the different steric hindrance effect of the thiol ligands directly results in the formation of Pt1Ag28 nanoclusters in different patterns.


In summary, two thiolated Pt1Ag28 nanoclusters, Pt1Ag28(S-2-Adm)20 and Pt1Ag28(S-2-Adm)18(HO-S-2-Adm)2, have been synthesized and characterized by SC-XRD and ESI-MS. The valence state of two S atoms of the latter nanocluster that bound to hydroxyls was found to be “−1”—the first time a sulfur atom bearing a “−1” valence state has been observed. SC-XRD analysis revealed that these nanoclusters crystallized in a 1:1 ratio and their structures forms the same icosahedral Pt1Ag12 kernel but different shell structures. ESI-MS analysis verified their cocrystallization and the presence of the hydroxyl on the nanocluster surface. Overall, these observations have significant implications for research into both metal nanoclusters and valence self-regulation, and further expand the chemistry of sulfur.



All reagents were purchased from Acros Organics or Sigma-Aldrich and used without further purification: hexachloroplatinic(IV) acid (H2PtCl6·6H2O, 99%, metals basis), silver nitrate (AgNO3, 99% metals basis), adamantane-1-thiol (C10H15SH, 95%), cyclohexyl mercaptan (HS-c-C6H11, 95%), 2-bromoadamantane (C10H15Br, 99.9%), sodium borohydride (NaBH4, 99.9%), sodium borodeuteride (NaBD4, 99.5%), potassium thioacetate (C2H3KOS, 99.9%), cesium acetate (CH3COOCs, 99% metals basis), methylene chloride [CH2Cl2, high-performance liquid chromatography (HPLC), Sigma-Aldrich], methanol (CH3OH, HPLC, Sigma-Aldrich), methanol-OD (CH3OD, 99.5 atom% D, Sigma-Aldrich), water-D2 (D2O, 99.9 atom% D, Sigma-Aldrich), water-18O (H218O, 98 atom% 18O, Sigma-Aldrich), ethanol (CH3CH2OH, HPLC, Sigma-Aldrich), ether (C2H5OC2H5, HPLC, Sigma-Aldrich), N,N-dimethylformamide (C3H7NO, HPLC, Sigma-Aldrich), and n-hexane (C6H14, HPLC, Sigma-Aldrich).

Synthesis of 2-AdmSH

2-AdmSH was prepared by the reported procedure (47). Mass spectral features with assigned fragments and relative intensities were observed at m/z of 168.08 (C10H15SH+), 134.73 (C10H15+), 93.11 (C7H9+), 91.07 (C7H7+), 79.09 (C6H7+), 77.04 (C6H5+), and 66.99 (C5H7+), in agreement with the mass features in (47).

Synthesis of Pt1Ag28(S-2-Adm)20 and Pt1Ag28(S-2-Adm)18(HO-2-S-Adm)2

In a 50-ml round-bottom flask, 20 mg of PPh3 was added to 20 ml of methylene chloride, and then a solution of 34 mg of AgNO3 as well as 5 mg of H2PtCl6·6H2O in 5 ml of methanol was added to the reaction, which resulted in a turbid yellow solution. After 20 min, a solution of 200 mg of 2-AdmSH in 2 ml of methylene chloride was added, and the solution turned clear. In this process, the metal-PPh3-thiol complexes were obtained. Shortly after this, a fresh solution of 20 mg of NaBH4 in 2 ml of ice-cold water was quickly added in. The color of the obtained solution turned brown immediately and then turned into reddish brown in 12 hours, indicating the formation of new nanoclusters. The obtained solution was centrifuged at 10,000 rpm for 5 min, and then the supernatant was collected and evaporated to get the dry products, which was then washed several times with ethanol to get the final products (13.4 mg, 28% yields calculated from AgNO3 to obtained nanoclusters).

Synthesis of Pt1Ag28(S-2-Adm)20 and Pt1Ag28(S-2-Adm)18(DO-2-S-Adm)2

The H2O, CH3OH, and NaBH4 reagents used to synthesize Pt1Ag28 nanoclusters (Pt1Ag28−1 and Pt1Ag28−2) were substituted by D2O, CD3OD, and NaBD4, respectively. Other conditions remained unchanged.

Synthesis of Pt1Ag28(S-2-Adm)20 and Pt1Ag28(S-2-Adm)18(H18O-2-S-Adm)2

The H2O used to synthesize Pt1Ag28 nanoclusters (Pt1Ag28−1 and Pt1Ag28−2) was substituted by H218O. Besides, 20 mg of NaBH4 was dissolved in 5 ml of ice-cold H218O to eliminate the interference of H2O from H2PtCl6·6H2O. Other conditions remained unchanged.

Single-crystal growth of Pt1Ag28(S-2-Adm)20 and Pt1Ag28(S-2-Adm)18(HO-2-S-Adm)2

Nanoclusters were crystallized in a CH2Cl2/ether system with a vapor diffusion method. Specifically, 20 mg of clusters was dissolved in 5 ml of CH2Cl2, and the obtained solution was then vapor-diffused by 50 ml of ether. After 3 days, dark crystals were collected and subjected to XRD to determine the structure.

Syntheses of Pt1Ag28(S-1-Adm)18(PPh3)4 or Pt1Ag28(S-c-C6H11)18(PPh3)4

The Pt1Ag28(S-1-Adm)18(PPh3)4 [or Pt1Ag28(S-c-C6H11)18(PPh3)4] nanocluster was prepared by the same method as the Pt1Ag28(S-2-Adm)20 and Pt1Ag28(S-2-Adm)18(HO-2-S-Adm)2 nanoclusters. The amount of 1-AdmSH (or HS-c-C6H11) was the same as that of 2-AdmSH.


All UV-vis absorption spectra of the nanoclusters that dissolved in CH2Cl2 were recorded using an Agilent 8453 diode array spectrophotometer. TGA was carried out using a thermogravimetric analyzer (DTG-60H, Shimadzu Instruments Inc.). XPS measurements were performed on a Thermo ESCALAB 250 instrument configured with a monochromated Al Kα (1486.8 eV) 150-W x-ray source, 0.5 mm circular spot size, flood gun to counter charging effects, and analysis chamber base pressure lower than 1 × 10−9 mbar. ICP–atomic emission spectrometry measurements were performed on an Atomscan Advantage instrument made by Thermo Jarrell Ash Corporation (USA). NMR spectra were acquired using a Bruker 600 Avance III spectrometer equipped with a Bruker BBO multinuclear probe (Bruker BioSpin, Rheinstetten, Germany); 20 mg of clusters was dissolved in 500 μl of CD2Cl2 and then subjected to the NMR measurement. IR measurements were recorded on a Bruker VERTEX 80sv Fourier transform IR spectrometer. ESI-TOF-MS measurements were performed using a MicrOTOF-QIII high-resolution mass spectrometer; to prepare the ESI sample, the clusters were dissolved in CH2Cl2 (1 mg/ml) and diluted (1:2, v/v) by dry methanol containing 5 mM CH3COOCs to ionize clusters by forming the cluster-Cs+. EPR spectra were acquired using a Bruker EMX Plus 10/12 spectrometer equipped with an Oxford ESR910 Liquid Helium cryostat; to prepare the EPR sample, 5 mg of clusters was dissolved in 500 μl of CH2Cl2 and then subjected to the EPR measurement.

X-ray crystallography

The data collection for single-crystal XRD was carried out using a Bruker Smart APEX II CCD diffractometer under nitrogen flow at 160 K, using graphite-monochromatized Mo Kα radiation (λ = 0.71073 Å). Data reductions and absorption corrections were performed using the SAINT and SADABS programs, respectively. Detailed information can be found in table S3. The structure was solved by direct methods and refined with full-matrix least squares on F2 using the SHELXTL software package. All nonhydrogen atoms were refined anisotropically, and all the hydrogen atoms were set in geometrically calculated positions and refined isotropically using a riding model. The solvent was squeezed by platon.


Supplementary material for this article is available at

Fig. S1. UV-vis, TGA, XPS, IR, 1H NMR, and EPR results of nanoclusters.

Fig. S2. Packing mode of nanoclusters.

Fig. S3. ESI-MS details of Pt1Ag28(SR)20 and Pt1Ag28(SR)18(DO-SR)2.

Fig. S4. ESI-MS of Pt1Ag28(SR)20 and Pt1Ag28(SR)18(H18O-SR)2.

Fig. S5. ESI-MS of the synthetic procedure.

Fig. S6. NMR signals of nanoclusters.

Fig. S7. UV-vis and ESI-MS results of Pt1Ag28(SR)18(PPh3)4.

Table S1. Atom ratio of Pt and Ag in nanoclusters.

Table S2. Comparison of bond lengths in nanoclusters.

Table S3. Crystal data and structure refinement for nanoclusters.

This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.


Acknowledgments: We thank S. Lacombe (Université de Pau et des Pays de l’Adour) for generous suggestions. EPR was performed on the Steady High Magnetic Field Facilities, High Magnetic Field Laboratory, CAS. Funding: We acknowledge the financial support by NSFC (U1532141, 21631001, 21871001, and 21803001), the Ministry of Education, the Education Department of Anhui Province, and 211 Project of Anhui University. Author contributions: M.Z. and S.W. conceived and designed the experiments. X.K., F.X., and X.W. conducted syntheses and characterizations. All authors discussed, wrote, and approved the final version of the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. Additional data related to this paper may be requested from the authors.

Stay Connected to Science Advances

Navigate This Article